N

SBa

b

c

d

e

a

ARRAA

KACAS

1

bfbaaiitopao(a

0h

Colloids and Surfaces B: Biointerfaces 102 (2013) 412– 419

Contents lists available at SciVerse ScienceDirect

Colloids and Surfaces B: Biointerfaces

jou rna l h om epa g e: www.elsev ier .com/ locate /co lsur fb

on-cytotoxic antibacterial silver–coumarin complex doped sol–gel coatings

warna Jaiswala,b, Kunal Bhattacharyac, Maeve Sullivand, Maureen Walshd,ernadette S. Creavend, Fathima Laffire, Brendan Duffya,∗, Patrick McHaleb

Centre for Research in Engineering Surface Technology (CREST), FOCAS Institute, Dublin Institute of Technology, Dublin 8, IrelandSchool of Biological Sciences, Dublin Institute of Technology, Kevin Street, Dublin 8, IrelandNanolab Research Centre, FOCAS Institute, Dublin Institute of Technology, Dublin 8, IrelandCentre for Pharmaceutical R&D, School of Science, Institute of Technology, Tallaght, Dublin 24, IrelandMaterials & Surface Science Institute, University of Limerick, Dublin, Ireland

r t i c l e i n f o

rticle history:eceived 29 May 2012eceived in revised form 27 July 2012ccepted 31 July 2012vailable online 16 August 2012

eywords:ntibacterialytotoxicityg–coumarin complexes

a b s t r a c t

Microbial colonisation on clinical and industrial surfaces is currently of global concern and silanebased sol–gel coatings are being proposed as potential solutions. Sol–gels are chemically inert, sta-ble and homogeneous and can be designed to act as a reservoir for releasing antimicrobial agentsover extended time periods. In the present study, silver nitrate (AgN) and a series of silver coumarincomplexes based on coumarin-3-carboxylatosilver (AgC) and it is 6, 7 and 8 hydroxylated analogues(Ag6, Ag7, Ag8) were incorporated into sol–gel coatings. The comparative antibacterial activity of thecoatings was determined against meticillin resistant Staphylococcus aureus (MRSA) and multidrug resis-tance Enterobacter cloacae WT6. The percentage growth inhibitions were found in the range of 9.2(±2.7)–66.0 (±1.2)% at low silver loadings of 0.3% (w/w) with E. cloacae being the more susceptible. Results

urface coating showed that among the Ag coumarin complexes, the Ag8 doped coating had the highest antibiofilmproperty. XPS confirmed the presence of silver in the nanoparticulate state (Ag0) at the coating sur-face where it remained after 4 days of exposure to bacterial culture. Comparative cytotoxicity studiesrevealed that the Ag-complex coatings were less toxic than the AgN coating. Thus, it can be concludedthat a sol–gel matrix with Ag–coumarin complexes may provide non-toxic surfaces with antibacterialproperties.

© 2012 Elsevier B.V. All rights reserved.

. Introduction

A critical problem facing the biomaterials sector is how to retardacterial growth and prevent biofilm formation on susceptible sur-aces. Bacteria can adhere to surfaces, multiply and form a compactiofilm matrix, which protects the underlying bacteria from thection of antibiotics, antibacterial agents and host defence mech-nisms. If this occurs with biomedical devices, such as prostheticmplants and catheters, it can result in serious infection, leading tomplant failure [1,2]. The risk of this occurring can be reduced byhe biomedical devices having antimicrobial surfaces in the formf treatments or coatings which prevent microbe adhesion androliferation [3]. For such applications, coatings need to combinentibacterial efficacy and low toxicity to eukaryotic cells. Vari-

us strategies such as (a) ionised/bound surface modification [6],b) incorporated metals in the form of salts [7], complexes [8]nd nanoparticles [9,10] and (c) organic molecules [11] have been

∗ Corresponding author. Tel.: +353 1402 7964; fax: +353 1402 7941.E-mail address: brendan.duffy@dit.ie (B. Duffy).

927-7765/$ – see front matter © 2012 Elsevier B.V. All rights reserved.ttp://dx.doi.org/10.1016/j.colsurfb.2012.07.047

developed to deliver a coating that can inhibit or remove bacte-ria [4,5], The sol–gel process can be used to form nano structuredinorganic films (typically 200 nm–10 �m in overall thickness) thatare more resistant than metals to oxidation, corrosion, erosion andwear while also possessing good thermal and electrical properties[12]. The technology has gained popularity very recently for thepreparation of porous coatings at relatively low temperature withdesirable surface chemistries, chemical durability [13] and ther-mal stability [14]. One of the most useful opportunities for sol–gelsis the inclusion of antibacterial metals and carboxylated organiccoumarin molecules which could produce an effective biomaterialcoating.

A number of recent studies have highlighted naturally derivedand synthetic coumarin (2H-1-benzopyran-2-one) derivatives withpotential antimicrobial [15,16] and antifungal activities [17]. Thereare a number of commercially available coumarin-based antibi-otics such as novobiocin, clorobiocin and coumermycin A1. It

has been reported that coordination of metal ions to therapeu-tic agents improves their efficacy and accelerates bioactivity.Coumarin derivatives can yield a wide variety of metal complexeswith different coordination modes, spectroscopic properties and

es B: B

plowbcl

ppsweetm4scdk

2

2

Tme

2

i5d2

fssl2

saxxrdp

2

1magafa

S. Jaiswal et al. / Colloids and Surfac

otential applications. In particular, coumarin-derived carboxylateigands can coordinate to various d-block metal ions in a numberf ways. Copper (II) complexes have binuclear bridged structureshile silver (I) complexes are also binuclear however the car-

oxylate group has a unidentate binding mode. Silver (I) coumarinomplexes possess excellent antimicrobial activity even though theigands themselves are inactive [15,18].

In this paper, silver nitrate and carboxylated coumarin com-lexes containing Ag were incorporated into sol–gel matrices toroduce homogeneous coatings in microtiter wells and on glasslides. These were analysed for biological activity and comparedith blank control equivalents. The chemical and biological prop-

rties of the coating surfaces were analysed. The coatings werexamined by scanning electron microscopy (SEM) after exposureo biofilm forming Staphylococcus epidermidis CSF 41498. Further-

ore, the chemical composition of the coatings before and after days bacterial exposure was determined by X-ray photoelectronpectroscopy (XPS). The release rate of silver ions (Ag+) from theoatings was studied and correlated with the antibacterial efficacyata. Finally, the cytotoxicity of the various coatings for humaneratinocyte cells was determined.

. Experimental

.1. Complex synthesis

The complexes were prepared as described previously [15,18].heir structure and purity were confirmed by thin-layer chro-atography, infrared analysis, 1H and 13C NMR spectroscopy and

lemental analysis.

.2. Preparation of sol–gels

All chemicals were purchased from Aldrich (Ireland).(a) Stock (blank) sol–gel: A blank, additive-free sol–gel, serv-

ng as a control, was prepared by mixing MTEOS with ethanol for min followed by the gradual addition of nitric acid and water asescribed previously [19]. The mixture was stirred continuously for4 h.

(b) Ag-doped sol–gels: Different levels of a dried silver salt andour silver–coumarin complexes were doped into aliquots of stockol–gel to give doping concentrations of 0.7, 0.5 and 0.3% (w/w) ofilver with respect to the final coatings weight (allowing for solventoss and curing). The modified sol–gel solutions were stirred for4 h in dark conditions to complete the formulation.

The silver salt gel (AgN) was made with silver nitrate and theol–gel. The complex doped sols were denoted as AgC, Ag6, Ag7nd Ag8 representing coumarin-3-carboxylatosilver (I), 6-hydro-ycoumarin-3-carboxylatosilver (I), 7-hydroxycoumarin-3-carbo-ylatosilver (I) and 8-hydroxycoumarin-3-carboxylatosilver(I)espectively. The complexes required the additional use of 5%imethyl sulfoxide (DMSO) in ethanol, as a solvent for solubilityurposes.

.3. Coating procedure

The sol–gels were applied as coatings on microtitre wells and cm × 1 cm glass slides. Sol–gels (200 �l) were dispensed into aicrotiter plate (Nunc, Denmark) and cured in a vacuum tight oven

t the 70 ◦C for 3 days [19]. The sol–gel were spin coated onto the

lass slides (specially coating instrument systems-spin coat G3P-8)t 1000 rpm with 5 s ramp and 1 min dwell time and cured at 70 ◦Cor 3 days. All coatings were stored at 4 ◦C prior to biofilm studiesnd XPS analysis.

iointerfaces 102 (2013) 412– 419 413

2.4. Antibacterial activity assay

The test bacteria, chosen multidrug resistant (ampicillin, gen-tamicin and Ceftazidime) Enterobacter cloacae WT6, meticillinresistant S. aureus (MRSA) ATCC 43300 and S. epidermidis CSF 41498were clinically significant antibiotic resistant strains. They weregrown, sub-cultured and maintained on Mueller-Hinton agar (LABM) and stored at 4 ◦C. For the experiment, a single colony of eachorganism was inoculated into Mueller-Hinton broth (MHB, 10 ml)and incubated overnight (24 h) at 37 ◦C with shaking at 200 rpm.The OD600 of the culture was adjusted to 0.132 (corresponding to1.5 × 108 CFU ml−1) using sterile MHB and further diluted to give afinal working concentration of 1.5 × 106 CFU ml−1.

The antibacterial activity of the Ag-doped sol–gel coatingsagainst the test bacteria was determined using the microtiter platemethod. Test bacteria (100 �l) from the 106 CFU ml−1 suspensionswere added to all test wells. A blank well containing the respec-tive cured Ag-doped sol–gel coating with sterile MHB (100 �l) anda negative control well containing the blank sol–gel with the bac-terial suspension (100 �l) were used. The microtiter plates wereincubated for 24 h at 37 ◦C.

The antibacterial activities of the different coatings were deter-mined by calculating the percentage inhibition of growth [20].The coating with the lowest incorporated metal concentrationthat completely inhibited bacterial growth after 24 h was consid-ered the minimum incorporated inhibitory concentration (MIIC).The minimum incorporated bactericidal concentration (MIBC) wasdetermined by the modified imprint method where the wellcontents (10 �l) were sub cultured onto MHA plates. The MIBCwas considered to be the lowest concentration that produced novisible bacterial growth on the MHA plate after 24 h at 37 ◦C indi-cating 99.5% killing of the original inoculum. All experiments wereperformed in triplicate and repeated at least twice.

2.5. Mathematical modelling of bacterial growth

2.5.1. Relation between turbidity and viable countIn order to convert the OD to log CFU ml−1, a relationship

between OD600 nm and viable count for the test bacteria wasdetermined [21]. An aliquot (200 �l) of bacterial suspension(1.5 × 106 CFU ml−1) was dispensed into the wells of the 96-wellcoated microtiter plate. Every hour, the OD of the well contents wasread and an aliquot (100 �l) from each of 3 wells was transferredinto maximum recovery diluent (MRD, 900 �l). This was furthertenfold diluted, and dilutions (100 �l) were plated out on MHAplates and incubated at 37 ◦C for 24 h, in order to determine thenumber of CFUml−1. A standard curve (OD600 nm vs. log CFU ml−1)was drawn from the results obtained.

2.5.2. Growth curve kineticsThe turbidity OD was measured at 600 nm at time intervals of

30 min, with 20 s agitation before each measurement, on a Power-wave microplate spectrophotometer (Powerwave, Biotek) drivenby Gen5 reader control and data analysis software. The OD valueswere converted to log CFU ml−1 by the standard curve as describedin Section 2.5.1.

Predictive modelling can be used to describe the growth curvesof the microorganisms over finite time periods. The growth charac-teristics can be determined using established iterative techniquesusing a least squares method. In general bacterial growth oftenshows a phase where specific growth rate starts at a value of zerofor a time period called a lag time (�). Thereafter growth acceler-

ates exponentially at a maximum rate (�max) before slowing andstopping (stationary period) so that a maximum colony count orasymptote (A) is reached. The value A therefore represents the max-imum colony count achievable under the experimental conditions.

4 es B: B

Tcectmsa

tolGV

l

wAvC

2

pbcttT1itfetmdtfr

2

cp1toan

2

urauaGt

14 S. Jaiswal et al. / Colloids and Surfac

he characteristic shape of such curves become sigmoidal whenolony counts are expressed in logarithmic terms. There are sev-ral models, which have been used to describe the sigmoid growthurve of a microorganism such as the Baranyi, Gompertz or Logis-ic models [22]. In general the Gompertz model is regarded as the

ost suitable to describe such microbial growth curves due to itsimplicity and the low correlation, or interdependence, of the char-cteristic parameters.

For the purposes of this work the modified Gompertz equa-ion [22] was fitted to the logarithm of the cell concentration inrder to estimate the maximum specific growth rates (�max) andag phase (�) of E. cloacae WT6 and MRSA ATCC 43300 using STAT-RAPHICS Centurion XV (StatPoint Technologies, Inc., Warrenton,A) statistical software. The model is described by Eq. (1):

og(

N

N0

)= A × exp

{− exp

[� × e

A(� − t) + 1

]}(1)

here � is the maximum specific growth rate of cell population, is the log increase of the population and � is the delayed timeariable (lag phase), N is the CFU ml−1 at any time t, N0 is the initialFU ml−1.

.6. Biofilm inhibition assay

S. epidermidis CSF 41498 (500 �l of 1.5 × 106 CFU ml−1) wasipetted into the coated microtiter wells (24 well plate) and incu-ated at 37 ◦C for 4 days. The control wells consisted of MTEOSoating with either sterile or bacterial loaded MHB. After incuba-ion, the suspensions were removed and the wells gently rinsedhrice with sterile water to remove loosely associated bacteria.he biofilms retained on the coating surfaces were stained with% (w/v) crystal violet solution (500 �l) for 45 min. After stain-

ng, coatings were washed thrice with sterile distilled water. Athis point, biofilms were visible as purple rings formed on the sur-ace of each well. Biofilm production was analysed by adding 95%thanol (500 �l) for 10 min to destain the wells. Aliquots (100 �l) ofhese were transferred to wells of another microtiter plate and OD

easured at 570 nm. The absorbance values were considered to beirectly proportional to the amount of biofilm biomass retained onhe coating surface. The results obtained were compared to thoseor an antimicrobial-free control coating to determine the relativeeduction in biofilm growth.

.7. Microscopic biofilm assay

Coated glass slides (1 cm2) spin coated with Ag8 sol–gel, ofoncentrations 0.5 and 0.7% (w/w), and the blank sol–gel werelaced in sterile petri dishes and S. epidermidis CSF 41498 (5 ml,.5 × 106 CFU ml−1) was added. After incubation at 37 ◦C for 4 days,he slides were removed, rinsed gently with sterile water, mountedn a stub and gold coated (Cressington 208HR sputter coater) fornalysis of the bacterial morphology using a Hitachi SU6600 scan-ing electron microscope (SEM) operating at 15 keV.

.8. Quantitative elementary analysis

The surface chemistry of the coatings was analysed by XPSsing a Kratos AXIS 165 spectrometer with monochromatic Al K�adiation of energy 1486.6 eV. High resolution spectra were takent fixed pass energy of 20 eV. Binding energies were determined

sing C 1s peak at 284.8 eV as charge reference. For constructionnd fitting of synthetic peaks of high resolution spectra, a mixedaussian-Lorenzian function with a Shirley type background sub-

raction were used.

iointerfaces 102 (2013) 412– 419

2.9. Evaluation of silver ion release rate

The release rate of the silver ions from the sol–gel coatingsinto sterile de-ionised water was analysed by ICP-AES (Varian Lib-erty 150). Plastic wells were attached to the surface of Ag8 andAgN sol–gel coated glass slides (0.7%, w/w) using an epoxy fixative(Araldite, Huntsman, UK) and left for 24 h [23]. Water (5 ml) wasadded to the wells (with a surface area of 4.9 cm2) and the slideswere incubated at 37 ◦C. Samples were removed and replaced withfresh water aliquots after 1, 4, 8, 12, 24, 48, 72 and 96 h. All leachateswere stored at 4 ◦C prior to pooling and acidification with nitric acid(0.1 M) for analysis as a single batch by ICP-AES, using a certifiedstock silver standard solution (ARISTAR® silver, BDH, England) forcalibration purposes.

2.10. In vitro cell culture

Immortalized non-cancerous human keratinocyte (HaCat) cellswere cultured in Dulbecco’s Modified Eagle’s Medium NutrientMixture F-12 HAM(DMEM-F12) supplemented with 10% foetalbovine serum (FBS) and antibiotic supplement (45 IU ml−1 peni-cillin and 45 IU ml−1 streptomycin) at 37 ◦C in 5% CO2.

2.11. Cytotoxicity assay

The coated microtiter wells (24 wells) were sterilized by UVradiation for 2 h. HaCat cells (500 �l of 1 × 105 cells ml−1 and8 × 104 cells ml−1) were added to the wells and incubated at37 ◦C for (24 and 48 h) in 5% CO2. MTEOS coated wells withHaCat cells were used as negative controls whereas cells withcis-diamminedichloroplatinum(II) (13 �M) served as a positivecontrol. Following the completion of the exposure time, an Ala-mar Blue (AB) assay was performed according to literature [24].Briefly, wells were rinsed with phosphate buffered saline (PBS)and 500 �l of AB medium (5%, v/v, in fresh culture medium with-out fetal bovine serum (FBS) or supplements) were added. After3 h of incubation, the AB fluorescence was measured at the excita-tion and emission wavelengths of 540 and 595 nm respectively, ina microplate reader (TECAN GENios, Grodig, Austria). Six replicatewells were used for each control and test, per microtiter plate, andthe experiment was repeated at least twice. The percentage cellviability was determined for all test coatings.

2.12. Statistical analysis

All the statistical analyses were done using STATGRAPHICSCenturion XV. The statistical significance (95% confidence level(P < 0.05)) between the means for the different Ag–coumarin com-plexes was determined by ANOVA followed by multiple range tests(least significant difference). Where inter-sample variation was notsignificant a group (denoted as a, b or c etc.) was defined. Withineach concentration, all samples denoted as “a” would be equivalentbut statistically different from samples denoted “b” and so on.

3. Results and discussion

The sol–gels (both blank and Ag doped) formed clear uni-form coatings in the microtitre wells after curing at the optimisedtemperature. Following curing the AgN coatings remained clearwhile the Ag–coumarin doped coatings developed a slight yel-

low colouring, characteristic of silver nanoparticles (Ag0). It maybe hypothesized that the Ag–coumarin complexes form weakbonds with the MTEOS at the low pH of 3–4, and may dissociateinto silver ion (Ag+) and coumarin (Cca). The curing temperature

S. Jaiswal et al. / Colloids and Surfaces B: Biointerfaces 102 (2013) 412– 419 415

Table 1Percentage growth inhibition of MDR E. cloacae WT6 and MRSA ATCC43300 bacteriaon the test sol–gel coatings with doping levels of 0.5% and 0.3% (w/w).

Samples Percentage (%) growth inhibition

Enterobacter cloacae WT6 MRSA ATCC 43300

0.3% 0.5% 0.3% 0.5%

Cca 2.94 ± 1.3a 4.3 ± 3.1a 1.2 ± 0.0a 12.26 ± 1.3a

AgC 9.2 ± 2.7ab 95.5 ± 8.7b 10.40 ± 1.3b 93.5 ± 7.8b

Ag6 17.9 ± 1.3bc 100 ± 0.0b 19.80 ± 4.0c 94.2 ± 9.7b

Ag7 25.6 ± 0.1c 97.5 ± 6.1b 31.60 ± 4.9d 98.4 ± 3.8b

Ag8 42.1 ± 4.3d 100 ± 0.0b 46.20 ± 1.3e 96.7 ± 4.4b

AgN 47.9 ± 7.1d 89.3 ± 2.3c 66.0 ± 1.2f 99.5 ± 5.9b

T(

pi

3

b0o(wptthcca

3

stgGrfed

pied(Aitiescrwh

i2w

5.2

5.7

6.2

6.7

7.2

7.7

8.2

8.7

0 5 10 15 20 25

LogC

FU

/ml

Time(h)

b

5.2

5.7

6.2

6.7

7.2

7.7

8.2

0 5 10 15 20 25

Log C

FU

/ml

Time (h)

Control

AgC

Ag6

Ag7

AgN

Ag8

a

Control

AgC

Ag6

Ag7

Ag8

AgN

face colonisation can be inhibited either by direct surface coatingsor with entrapped agents that are gradually released. Such hygieniccoatings can be employed on environmental surfaces and medicaldevices. The percentage inhibition of S. epidermidis biofilm growth

Table 2Value of model parameters [Lag phase (�) and maximum specific growth rate (�max)]for the non-linear regression of Gompertz equation for fitting of bacterial growth.

� (h) �max (h−1) R2

E. cloacae WT6Control 5.89 ± 0.04 0.32 ± 0.00 0.996AgC 6.96 ± 0.04 0.19 ± 0.01 0.998Ag6 8.66 ± 1.22 0.17 ± 0.03 0.994Ag7 9.14 ± 0.31 0.15 ± 0.01 0.996Ag8 9.70 ± 0.45 0.14 ± 0.02 0.997AgN 10.26 ± 0.49 0.12 ± 0.02 0.991

MRSA ATCC 43300Control 2.79 ± 1.22 0.22 ± 0.02 0.990AgC 4.88 ± 0.76 0.18 ± 0.05 0.990

he letters (a–f) at each concentration indicate groups that are significantly differentP < 0.05)

romoted the development of the yellowing indicating that theres a thermodynamic process occurring.

.1. Antibacterial activity

The percentage growth inhibition of E. cloacae WT6 and MRSAy the test sol–gel coatings containing varying doping levels (0.3%,.5% and 0.7%, w/w) was determined. There was a complete absencef bacterial growth for all Ag based coatings at a loading level of 0.7%w/w). At 0.5% (w/w) loadings significant levels of inhibition (>89%)ere observed in all Ag coatings (Table 1), whereas at 0.3% (w/w)artial inhibition of bacterial growth was observed (critical concen-ration). The MIICs and MIBCs of all Ag doped coatings were foundo be 0.5% (w/w) against both test bacteria. Results showed that Ccaad no significant antibacterial activity. Among the Ag–coumarinomplexes, Ag8 coatings showed significantly (P < 0.05) higher per-entage growth inhibition at the critical concentration 0.3% (w/w)lthough somewhat lower than the base reference AgN.

.2. Growth curve kinetics

Traditional microbial enumeration techniques are time con-uming and therefore, mathematical models are regularly usedo establish the kinetic parameters of survival curves of microor-anisms. The empirical sigmoid type models such as the modifiedompertz, Logistic or Baranyi models have been used to fit bacte-

ial growth [25], however, these models are usually only applied toood processing and food safety studies. As per the author knowl-dge this is first time modeling has been used for the analysis ofelaying or inhibition of bacterial growth on the coating surfaces.

Different concentrations of incorporated Ag–coumarin com-lexes showed variable levels of bacterial inhibition over 24 h of

ncubation (Fig. 1). The coatings showed a strong antagonizingffect on the test bacteria studied, displaying a significant dose-ependent relationship with an increase in the lag phase duration�) and corresponding decrease in the exponential growth rate (�).

modified Gompertz model was applied to estimate the delay ornhibition in bacterial (MRSA and E. cloacae) growth. In most cases,he correlation (R2) values for the models were greater than 99%ndicating a good fit to the experimental data (Table 2). It was alsostablished that the maximum specific growth rate (�max), wasignificantly (P < 0.05) reduced in the presence of the Ag dopedoatings. Reduction in the �max for the Ag–coumarin coatingsanged from 41 to 56% for E. cloacae and 18 to 41% for MRSAith respect to control. The corresponding reductions for AgN wereigher at 62% and 95% respectively.

At the critical concentration (0.3%) the Ag–coumarin coatingsncreased the lag phase from 1 to 3.81 h for E. cloacae and from.09 to 5.35 h for MRSA. Among the Ag–coumarin coatings Ag8as the most efficient, increasing the lag phase of E. cloacae and

Fig. 1. Fitting of the model to the inactivation of the bacteria by the coatings contain-ing various Ag-complexes (control, AgC, Ag6, Ag7, Ag8 and AgN) at a concentrationof 0.3% (w/w) over 24 h (a) E. cloacae WT 6 and (b) MRSA.

MRSA by 64.68% and 191.75% with respect to control. It is hypothe-sized that small structural differences in the Ag-complexes causedsome changes in the biological activity. The position of the hydroxylgroup in these Ag-complexes is known to contribute significantlyto antimicrobial activity [17].

3.3. Biofilm inhibition assay

Biofilm formation is a multi-factorial process involving theadhesion of bacterial cells to a host substrate, followed by subse-quent attachment and colonisation [23]. S. epidermidis is the mostcommon bacterium isolated from medical devices such as vascularcatheters, prosthetic implants and intrauterine devices [26]. Sur-

Ag6 6.28 ± 0.10 0.14 ± 0.01 0.993Ag7 7.41 ± 1.44 0.13 ± 0.01 0.986Ag8 8.14 ± 0.08 0.13 ± 0.00 0.984AgN 7.79 ± 1.29 0.01 ± 0.01 0.994

416 S. Jaiswal et al. / Colloids and Surfaces B: B

Fig. 2. Percentage inhibition of biofilm forming S. epidermidis CSF41498 on sol–gelcoatings ( ) Ag6, ( ) Ag7, ( ) Ag8, ( ) AgC and ( ) AgN at concentrations0a

otws

wttA0At2

.3%, 0.5% and 0.7% (w/w), after 4 days incubation. *Letter (a–e) in each concentrationre significantly different (P < 0.05) for various Ag–coumarin complexes and AgN.

n Ag doped sol–gel coatings is displayed in Fig. 2. In comparison tohe blank control, a reduction in biofilm formation of about 25–80%as observed in the presence of Ag–coumarin coatings across the

tudied concentration range (0.3–0.7%, w/w).Among the Ag-doped coatings, maximum biofilm inhibition

as observed with Ag8 while AgC showed the least inhibi-ion. In terms of performance, there was a clear ranking amonghe Ag–coumarin sol–gel coatings at all concentrations, namelyg8 > Ag7 > Ag6 > AgC, with the last two position reversed at

.3% (w/w). Interestingly the improvement in inhibition of theg–coumarin coatings with concentration was not as clear as with

he AgN coatings. AgN coatings showed antibiofilm activity from1% at 0.3% (w/w) up to 73% at 0.7%. It was also evident from the

Fig. 3. SEM images of S. epidermidis CSF 41498 on (a) blank control coated glass s

iointerfaces 102 (2013) 412– 419

results that there was significant (P > 0.05) difference between Ag8,Ag7 and AgN coatings at 0.3% and 0.5% whereas; at 0.7% AgN therewas not.

3.4. Microscopic biofilm assay

The effect of silver ion release from the Ag8 coatings was stud-ied by SEM (Fig. 3). After 4 days of incubation, the biofilm was fullyestablished on the blank control coating and the glass slide wascovered with adherent S. epidermidis CSF 41498 (Fig. 3a). Silverion release from the doped Ag8 reduced the bacterial colonisa-tion on the coating surface and distinctive crystalline precipitateswere seen (Fig. 3b and c). These may be due to insoluble adher-ent salt crystals on the coating surface. Extraneous matter such aschloride from the culture medium may have coupled with some ofthe released silver ions to form silver chloride at the surface in thebroth.

The Ag8 coating greatly reduced biofilm formation at 0.5%(Fig. 3b) and produced a higher number of damaged bacterial cellsat 0.7% (Fig. 3c). These results were in agreement with some pre-vious work, which showed that silver ions act at the peripheralarea of S. epidermidis biofilm causing disruptive changes that desta-bilize it [27]. Moreover, rapid elution was found to be critical assilver ions become inactive after binding with protein anions andpolysaccharides in the biofilm.

3.5. Quantitative elementary analysis

To describe the antibacterial activity of a coating, an understand-ing of the surface properties and change in its chemical state are

important. XPS was used to study the change in the surface chem-ical state of the most active coating (Ag8) and the reference (AgN)before and after immersion for 4 days in an S. epidermidis CSF 41498culture. From the summary of the elemental composition given in

lide, (b) Ag8 0.5% (w/w) coated slide and (c) Ag8 0.7% (w/w) coated slides.

S. Jaiswal et al. / Colloids and Surfaces B: Biointerfaces 102 (2013) 412– 419 417

Table 3Summary of XPS data, before and after exposure to Staphylococcus epidermidis broth for four days.

at.%

Coating type C O Si Ag N Na

C C C N/C O N C O/O C O

AgN 28.3 43.1 27.8 0.8 0.0 0.0(27.4) (0.9) (0)

Ag8 28.5 45.1 26.3 0.1 0.0 0.0(27.9) (0.6) (0)

AgN-after exposure 46.0 31.4 14.0 0.2 6.7 1.6

32.0 16.2 0.1 5.0 0.4

Twi2tCri6ttceti

e3aaaiic

cTtss

0

30

60

90

120

150

968880726456484032241680

Con

cen

tra

tion

(p

pb

)

Time (h)

Ag8

AgN

FS

(28.1) (11.3) (6.6)Ag8-after exposure 46.3

(31.9) (9.2) (5.2)

able 3 a significant increase in the relative carbon concentrationsas seen after immersion in both cases. Before exposure carbon

s present predominantly as hydrocarbon at a binding energy of84.8 eV. However, after exposure the increase in carbon con-ent is attributed to additional presence of carbonaceous species,

O/C N at ∼286.5 eV and O C O/N C O at ∼288.5 eV. Nitrogenelated to these functional groups appears in the coatings at a bind-ng energy of 400 eV after exposure in significant amounts (7.8 and.0 at.%). These are likely to originate from proteinaceous species inhe broth, assuming the damaged bacterial cell wall released pro-ein into the broth. In addition, after exposure to the broth bothoatings showed the presence of a small amount of sodium. How-ver, the relative concentration of Na in Ag8 is notably less (0.4%)han in AgN (1.6%) perhaps indicating interaction of the coumarinon with Na+.

High resolution Ag 3d spectra of the Ag8 coating before and afterxposure to S. epidermidis CSF 41498 are shown in Fig. 4a. The Agd peak appears as a doublet of Ag 3d5/2 and Ag 3d3/2 at 368.1 eVnd 374.5 eV respectively with a doublet separation of 6.0 eV char-cteristic of metallic Ag. After exposure, the Ag 3d5/2 peak shifts to

lower binding energy of 367.9 eV corresponding to oxidised Ag+

ndicating that the Ag0 has been released in the form of Ag+. A sim-lar trend in the binding energies was observed for the AgN dopedoating (Fig. 4b).

The relative composition of silver present in the Ag8 and AgNoatings was 0.1% and 0.8% pre bacterial incubation (Table 3).

he silver content decreased significantly post incubation inhe AgN coating, whereas in the Ag8 coating it showed amall change. This may suggest a continuous slow release ofilver ion from Ag8 coating over the extended time period

364366368370372374376378380

Binding energy (eV)

Intensity

(a.u.)

Intensity

(a.u.)

(a)

ig. 4. High resolution Ag 3d XPS spectra of (a) AgN doped coating and (b) Ag8 doped coat. epidermidis CSF 41498.

Fig. 5. Non-cumulative release rates of Ag into fresh DI H2O from Ag8 and AgNcoated glass slides as a function of time (determined by ICP emission spectroscopy).

compared to AgN, which has a high initial silver release rate[28].

3.6. Evaluation of silver ion release rate

In order to possess effectiveness against microorganisms, sil-ver ions should be capable of migrating from the hybrid matrix of

the coatings. Fig. 5 shows the release of silver from Ag8 and AgNcoatings containing 0.7% (w/w) in to de-ionised water as a functionof time at 37 ◦C. The release of silver from both the Ag8 and AgNcoatings was characterised by an initial high release followed by a

364366368370372374376378380

Binding energy (eV)

(b)

ing before (dark curve) and after (grey curve) 4 days of exposure to biofilm forming

418 S. Jaiswal et al. / Colloids and Surfaces B: B

Fig. 6. Cytotoxic response of HaCat cells after (a) 24 and (b) 48 h of exposure to Ag-doped sol–gel coatings (( ) Ag6, ( ) Ag7, ( ) Ag8, ( ) AgC and ( ) AgN) withincreasing concentration of 0.3%, 0.5% and 0.7% (w/w). *Letter (a–e) in each concen-tration are significantly different (P < 0.05) for various Ag–coumarin complexes andA

st>ttwbo

mtiahtdtmpa[

eaetmmAtedA

[6] M. Chen, L.Y. Wang, J.T. Han, J.Y. Zhang, Z.Y. Li, D.J. Qian, J. Phys. Chem. B 110(2006) 11224.

gN.

harp drop over the first 24 h. The release rate from AgN decreasedo less than 10 ppb after 12 h, whereas Ag8 maintained a release of20 ppb for 72 h, dropping to 13.9 ppb after 96 h. The results implyhat the solubility of silver in the AgN coating is higher than inhe coumarin complex. From a design perspective a silver coatingith a high release rate is not suitable as an antibacterial coating for

iomaterials because the antibacterial agent can be quickly washedut [29].

After 96 h, the controlled release rate of Ag+ from Ag8 is stillaintained at bactericidal levels. The release rate of Ag+ depends on

he concentration of nanoparticles and their diffusion to the coat-ng surface. Rapid diffusion of the broth solution into the coatingnd higher accumulation of silver at the surface could result in aigh metal release rate in the initial stages. Significant reduction inhe release rate of Ag+ in later stages can be explained by the rapidecrease of surface Ag particles and slower diffusion of the solu-ion in the pores of the coating, thereby changing the Ag+ release

echanism [30]. This diffusion release profile is consistent withorous systems and is in contrast to swelling materials which have

rapid ion release rate after a minimum release in the initial stage31,32].

The release of Ag+ from silver nanoparticles is primarily a het-rogeneous oxidation reaction involving oxygen chemisorptionccompanied by electron transfer [33]. Therefore, methods thatnhance or disrupt oxidation pathways are a promising routeo controlled release. Several routes have been explored using

acromolecules such as dextrans [20,34], starch [35], and poly-ers [36,37] which can block oxygen access. In this work, theg–coumarin coating serves as a source for Ag+ in solution and

he bound Ag+ undergoes reversible adsorption–desorption influ-+

nced by the Ag concentration. Coumarin coatings can both

elay and extend ion release by accumulating and releasingg.

iointerfaces 102 (2013) 412– 419

3.7. Cytotoxicity study

Cytotoxicity is a critical biocompatibility issue for synthesizedcompounds, including metal doped sol–gels. Published reportsconfirm that while Ag+ concentrations less than <35 ppb are bacte-ricidal [38] higher concentrations between 1 ppm and 10 ppm maycause damage to human cells. This damage can range from lowlocalised toxicity to conditions like argyria in higher doses (>4–6 gin the body) [39]. Thus it is important to evaluate the cytotoxicityof the synthesized metal doped sol–gels coatings.

The study showed a dose-dependent response for all Ag coatings(Fig. 6a and b), whereby the toxicity increased with increasing con-centration from 0.3% to 0.7% (w/w). Following 24 h exposure tothe Ag–coumarin coatings, the percentage cell viability was in theorder of Ag7 ≥ Ag8 > Ag6 > AgC > AgN at almost all tested concentra-tions (0.3%, 0.5% and 0.7%) (Fig. 6a). In the presence of Ag7 and Ag8coatings, cell viability was up to 70–80% at higher concentration(0.7%) with respect to negative control. After 48 h of incubation(Fig. 6b) the cell viability for all coatings fell by approx. 10–20%between 24 and 48 h. Importantly, the comparative cell viabilitywas significantly (P < 0.05) higher for the Ag–coumarin coatingsthan for the AgN coating at all concentrations.

4. Conclusion

In summary, Ag doped sol–gels deliver a porous coating aftercuring at 70 ◦C for 3 days. Antibacterial Ag–coumarins displayedactivity within coatings even at a low concentration of 0.3% (w/w)with a reproducible difference in activity between tested com-plexes. Of the four coumarin coatings tested against the biofilmforming bacterium Ag8 sol–gel was shown to have significantantibiofilm properties at concentrations of 0.5% and 0.7% (w/w).The Ag–coumarin coatings showed less toxicity than AgN equiv-alents. Among all coatings Ag7 had the lowest toxicity while Ag8had the highest antibacterial activity. XPS results confirmed thatthe silver was in its metallic form (Ag0) within the Ag coumarincoating. There was a gradual release of Ag+ from the Ag coumarincoating over an extended time period which maintained a higherefficacy. The coumarin reduced the toxicity of Ag in the coating;it is hypothesized that the prolonged release of Ag+ is facilitatedby the carboxylated coumarins, creating a less cytotoxic surface.These results provide evidence for the potential use of Ag–coumarinsol–gels coatings as biomedical coatings.

Acknowledgements

The authors would like to gratefully acknowledge Mr. AmitKumar Jaiswal for assistance in mathematical modelling and Ms.Anne Shanahan for SEM imaging. The authors would also like toacknowledge the Dublin Institute of Technology, Dublin, Irelandfor funding under the ABBEST scholarship program and the Centreof Applied Science and Health at ITT Dublin.

References

[1] V. Sambhy, M.M. MacBride, B.R. Peterson, A. Sen, J. Am. Chem. Soc. 128 (2006)9798.

[2] D. Zhang, O. Lepparanta, E. Munukka, H. Ylanen, M.K. Viljanen, E. Eerola, M.Hupa, L. Hupa, J. Biomed. Mater. Res. A 93 (2010) 475.

[3] C.H. Ho, J. Tobis, C. Sprich, R. Thomann, J.C. Tiller, Adv. Mater. 16 (2004) 957.[4] L. Zhao, P.K. Chu, Y. Zhang, Z. Wu, J. Biomed. Mater. Res. B: Appl. Biomater. 91

(2009) 470.[5] L. Zhang, D. Pornpattananangkul, C.M.J. Hu, C.M. Huang, Curr. Med. Chem. 17

(2010) 585.

[7] O. Choi, K.K. Deng, N. Kim, L. Ross, R.Y. Surampalli, Z. Hu, Water Res. 42 (2008)3066.

[8] Y. Li, W.K. Leung, K.L. Yeung, P.S. Lau, J.K.C. Kwan, Langmuir 25 (2009) 13472.

es B: B

[

[[

[

[

[

[

[

[

[[

[

[

[

[

[

[[

[

[[[

[[[

[

[

[[38] A. Ewald, S.K. Glückermann, R. Thull, U. Gbureck, Biomed. Eng. Online 5 (2006)

S. Jaiswal et al. / Colloids and Surfac

[9] D. Ionita, M. Grecu, C. Ungureanu, I. Demetrescu, Appl. Surf. Sci. 257 (2011)9164–9168.

10] A. Travan, C. Pelillo, I. Donati, E. Marsich, M. Benincasa, T. Scarpa, S. Semeraro,G. Turco, R. Gennaro, S. Paoletti, Biomacromolecules 10 (2009) 1429.

11] M. Vallet-Regí, M. Colilla, B. González, Chem. Soc. Rev. 40 (2011) 596.12] P.C.R. Varma, J. Colreavy, J. Cassidy, M. Oubaha, B. Duffy, C. McDonagh., Prog.

Org. Coat. 66 (2009) 406.13] P.C.R. Varma, J. Colreavy, J. Cassidy, M. Oubaha, C. McDonagh, B. Duffy, Thin

Solid Films 518 (2010) 5753.14] P.C.R. Varma, P. Periyat, M. Oubaha, C. McDonagh, B. Duffy, Surf. Coat. Technol.

205 (2011) 3992.15] B.S. Creaven, D.A. Egan, K. Kavanagh, M. McCann, A. Noble, B. Thati, M. Walsh,

Inorg. Chim. Acta 359 (2006) 3976.16] C. Gnerre, M. Catto, F. Leonetti, P. Weber, P.A. Carrupt, C. Altomare, A. Carotti,

B. Testa, J. Med. Chem. 43 (2000) 4747.17] B. Thati, A. Noble, R. Rowan, B.S. Creaven, M. Walsh, M. McCann, D. Egan, K.

Kavanagh, Toxicol. In Vitro 21 (2007) 801.18] B.S. Creaven, D.A. Egan, K. Kavanagh, M. McCann, M. Mahon, A. Noble, B. Thati,

M. Walsh, Polyhedron 24 (2005) 949.19] S. Jaiswal, P. McHale, B. Duffy, Colloids Surf. B: Biointerfaces 94 (2012) 170.20] S. Jaiswal, B. Duffy, A.K. Jaiswal, N. Stobie, P. McHale, Int. J. Antimicrob. Agents

36 (2010) 280.21] A.K. Jaiswal, S. Gupta, N. Abu-Ghannam, S. Cox, Food Sci. Technol. Int. 17 (2011)

495.22] M.H. Zwietering, I. Jongenburger, F.M. Rombouts, K. Van ‘t Riet, Appl. Environ.

Microbiol. 56 (1990) 1875.23] N. Stobie, B. Duffy, J. Colreavy, P. McHale, S.J. Hinder, D.E. McCormack, J. Colloid

Interface Sci. 345 (2010) 286.

[

iointerfaces 102 (2013) 412– 419 419

24] S.P. Mukherjee, F.M. Lyng, A. Garcia, M. Davoren, H.J. Byrne, Toxicol. Appl.Pharmacol. 248 (2010) 259.

25] S. Gupta, S. Cox, G. Rajauria, A.K. Jaiswal, N. Abu-Ghannam, Food BioprocessTechnol. 5 (2012) 1907.

26] P.S. Brunetto, K.M. Fromm, CHIMIA Int. J. Chem. 62 (2008) 249.27] K. Chaw, M. Manimaran, F.E.H. Tay, Antimicrob. Agents Chemother. 49 (2005)

4853.28] N. Stobie, B. Duffy, D.E. McCormack, J. Colreavy, M. Hidalgo, P. McHale, S.J.

Hinder, Biomaterials 29 (2008) 963.29] B. Mahltig, D. Fiedler, P. Simon, J. Text. Inst. 102 (2011) 739.30] R. Kumar, H. Münstedt, Biomaterials 26 (2005) 2081.31] A. Babapour, B. Yang, S. Bahang, W. Cao, Nanotechnology 22 (2011)

155602.32] C. Damm, H. Münstedt, A. Rösch, Mater. Chem. Phys. 108 (2008) 61.33] J. Liu, D.A. Sonshine, S. Shervani, R.H. Hurt, ACS Nano 11 (2010) 6903.34] Y.Q. Ma, J.Z. Yi, L.M. Zhang, J. Macromol. Sci. A: Pure Appl. Chem. 46 (2009)

643.35] P. AshaRani, G. Low Kah Mun, M.P. Hande, S. Valiyaveettil, ACS Nano 3 (2008)

279.36] X. Cai, M. Lin, S. Tan, W. Mai, Y. Zhang, Z. Liang, Z. Lin, X. Zhang, Carbon 50

(2012) 3407.37] R.B. Grubbs, Polym. Rev. 47 (2007) 197.

22.39] G. Gosheger, J. Hardes, H. Ahrens, A. Streitburger, H. Buerger, M. Erren, A.

Gunsel, F.H. Kemper, W. Winkelmann, C. von Eiff, Biomaterials 25 (2004)5547.